NetBSD Problem Report #55363
From john@ziaspace.com Wed Jun 10 22:02:30 2020
Return-Path: <john@ziaspace.com>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
by mollari.NetBSD.org (Postfix) with ESMTPS id 5DF6B1A9219
for <gnats-bugs@gnats.NetBSD.org>; Wed, 10 Jun 2020 22:02:30 +0000 (UTC)
Message-Id: <202006102202.05AM2Ngn004088@athena.zia.io>
Date: Wed, 10 Jun 2020 22:02:23 GMT
From: john@ziaspace.com
Reply-To: john@ziaspace.com
To: gnats-bugs@NetBSD.org
Subject: NetBSD/i386 PAE has issues
X-Send-Pr-Version: 3.95
>Number: 55363
>Category: port-i386
>Synopsis: NetBSD/i386 PAE has issues with memory and NVMe
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: port-i386-maintainer
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Wed Jun 10 22:05:00 +0000 2020
>Originator: John Klos
>Release: NetBSD 9.0_STABLE
>Organization:
>Environment:
System: NetBSD doozer.zia.io 9.0_STABLE NetBSD 9.0_STABLE (GENERIC_PAE) #0: Fri May 22 04:28:49 UTC 2020 john@helena.zia.io:/usr/obj-i386/sys/arch/i386/compile/GENERIC_PAE i386
Architecture: i386
Machine: i386
>Description:
NetBSD 9.0, NetBSD 9 as of 10-June-2020 has two noticeable problems:
On a Ryzen 3600 system with 64 gigs and a PAE kernel, only 3.5 gigs is usable:
[ 1.000000] total memory = 3527 MB
[ 1.000000] avail memory = 3445 MB
When memory is removed so the system has just 32 gigs, all 32 gigs is usable:
[ 1.000000] total memory = 32699 MB
[ 1.000000] avail memory = 32112 MB
With 64 gigs (3.5 usable), a PCIe NVMe card is usable as expected:
[ 1.046309] nvme0 at pci3 dev 0 function 0: vendor 8086 product 0953 (rev. 0x01)
[ 1.046309] nvme0: NVMe 1.0
[ 1.046309] nvme0: for admin queue interrupting at msix0 vec 0
[ 1.046309] nvme0: INTEL SSDPEDMW400G4, firmware 8EV101H0, serial CVCQ542200BT400AGN
[ 1.046309] nvme0: for io queue 1 interrupting at msix0 vec 1 affinity to cpu0
[ 1.046309] nvme0: for io queue 2 interrupting at msix0 vec 2 affinity to cpu1
[ 1.046309] nvme0: for io queue 3 interrupting at msix0 vec 3 affinity to cpu2
[ 1.046309] nvme0: for io queue 4 interrupting at msix0 vec 4 affinity to cpu3
[ 1.046309] nvme0: for io queue 5 interrupting at msix0 vec 5 affinity to cpu4
[ 1.046309] nvme0: for io queue 6 interrupting at msix0 vec 6 affinity to cpu5
[ 1.046309] nvme0: for io queue 7 interrupting at msix0 vec 7 affinity to cpu6
[ 1.046309] nvme0: for io queue 8 interrupting at msix0 vec 8 affinity to cpu7
[ 1.046309] nvme0: for io queue 9 interrupting at msix0 vec 9 affinity to cpu8
[ 1.046309] nvme0: for io queue 10 interrupting at msix0 vec 10 affinity to cpu9
[ 1.046309] nvme0: for io queue 11 interrupting at msix0 vec 11 affinity to cpu10
[ 1.046309] nvme0: for io queue 12 interrupting at msix0 vec 12 affinity to cpu11
[ 1.046309] ld0 at nvme0 nsid 1
However, with 32 gigs (32 usable), the NVMe card stops working:
[ 1.038776] nvme0 at pci3 dev 0 function 0: vendor 8086 product 0953 (rev. 0x01)
[ 1.038776] nvme0: NVMe 1.0
[ 1.038776] nvme0: for admin queue interrupting at msix0 vec 0
[ 1.038776] nvme0: INTEL SSDPEDMW400G4, firmware 8EV101H0, serial CVCQ542200BT400AGN
[ 1.038776] nvme0: autoconfiguration error: unable to allocate ccbs
[ 1.038776] nvme0: autoconfiguration error: unable to allocate io queue
>How-To-Repeat:
>Fix:
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.46 2020/01/03 16:35:01 leot Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2020
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.